Auto transcribe new voice memo in Drafts

I am looking for a way to use the transcribing feature in Drafts (Transcribing Audio & Video - Drafts User Guide) to automatically transcribe audio memos.

The dream is to have a process like:
Pressing record in Apple Voice Memos → making the recording → pressing stop → Drafts automaticly triggers and starts transcribing the audio.

Additionally it would be nice if the finished draft included a link to the audio file in iCloud.

I have tied playing around with it in Shortcuts, but can’t really find a way to trigger the transcribing feature. Wondering if it would be able to do something like this with a Drafts action triggered by Shortcuts?

Thought maybe one of you Shortcut and Drafts wizards would have some ideas to how this could be solved.

I poked around and couldn’t find any hooks into Drafts’ Transcribe feature or even feeding a file into iOS’s native Dictate text.

One option is to just dictate directly into Drafts. Use a Dictate Text action and feed it into a new or existing draft.

This looks a lot like your request on the Drafts forum from earlier this week, which I provided a few ideas for.

You can create a Drafts action with a scripting step that calls the editor.transcribe() function. That can then be triggered from Shortcuts (various methods, but I would probably opt for the Run action with text Shortcuts action that Drafts provides).

The issue is that it does not provide an option to select a file to process. It simply displays the transcription interface. That is where, as per the Drafts topic linked above, you would need to look at external tools to help you manipulate that. If you are restricted to i*OS, then you are out of luck. If you are on the Mac, then tools like Keyboard Maestro, or maybe even the UI interaction actions that Better Touch Tool provides to Shortcuts may be an option for you.

The other option, as noted in the linked Drafts topic, is to nudge Greg to see if he could provide a way to automate that feature. For example, adding an optional file path parameter to the command above might be viable.

As a side note, one other way I had considered (and tested) was that you can get Drafts to process audio files for transcription by dragging and dropping the file onto the app. For an mp3 file called rec.mp3 on your Desktop, yoiu can actually produce the same effect on the command line:

open - a "Drafts" ~/Desktop/rec.mp3

Unfortunately, whjile this gets you around the having to interact with the UI and choose a file, you get a similar issue to editor.transcribe() in that you are presented with UI elements that you need to interact with. One to initate the transcription, and one to commit the content to a new draft at the end.

That isn’t a native feature as Dictation is live input - and actually how several ‘cheap’ techniques for transcribing audio files work when you look it up online - effectively put your speaker next to your microphone :roll_eyes:

But, that is where an app like Actions for Shortcuts comes in (see the Drafts topic linked above), or where a developer builds a route into their app (ref. Drafts).

@havarha - based on the above, are there any more specifics around what you have tried/not tried, do/do not have access to, etc? There may even be other options you could consider such as utilising paid for web services to do the transcription, or even a Mechanical Turk (ref. platform) approach.

1 Like